An open tool for creating battery-electric vehicle time series from empirical data -- emobpy

• The tool’s first application creates vehicle profiles based on representative German mobility data.Source: §1 · p. 2
• Each profile contains time series for vehicle mobility, driving electricity consumption, grid availability, and grid electricity demand.Mobility records location and travel; grid availability records connection and charging or discharging power; demand records electricity drawn from the grid.Source: §1 · p. 2
• emobpy derives profiles from empirical mobility statistics, vehicle properties, and customizable assumptions.Source: §1 · p. 3
• Mobility profiles use trip and departure information, with Monte Carlo sampling to vary profiles and constraints to keep them feasible and consistent.The described constraints include minimum and maximum time at locations and a final daily trip home.Source: §1 · p. 4
• Driving electricity consumption is derived from mobility and vehicle inputs, including motor power, weight, and drag coefficient.The tool also provides temperature data and driving cycles, while allowing user-defined data.Source: §1 · p. 4
• Grid electricity demand uses the consumption and grid-availability series, with user-selected charging strategies such as immediate full or night-time charging.Source: §1 · p. 4

2.1 Application to Germany: parameterization and setup

• The Germany application generates 200 profiles—50 each for four BEV models—assuming future BEV drivers have mobility patterns like current conventional drivers.The application draws on the German Mobilit¨at in Deutschland survey; its assumptions can be modified in emobpy.Source: §2.1 · pp. 4–5
• Four illustrative charging strategies range from immediate full-power charging to immediate balanced, home-only balanced, and night-time home-balanced charging.The night-time strategy restricts charging to 23:00–8:00; none of the strategies accounts for power-sector conditions or electricity-market prices.Source: §2.1 · pp. 7–8

2.2 Vehicle mobility

• On working days, workplace presence peaks at 32% at 11:00, while the fleet’s driving share peaks around 15:00–17:00 at about 9%.Every day between 6:00 and 22:00, at least 3% of the fleet are driving.Source: §2.2 · p. 8
• Both emobpy and official German statistics indicate that about 90% of all trips have distances of 50 km or below, and cumulative mileage has a similar shape in the two sources.Cumulative mileage refers to the overall distance travelled by all vehicles in a year.Source: §2.2 · p. 8
• The mobility results associate the discussed trips with yearly mileage while reporting that they account for 3% of all trips.Source: §2.2 · p. 8

2.3 Driving electricity consumption

• For one million BEV, weekday driving electricity consumption peaks between 15:00 and 17:00 with an annual median around 450 MWh, reaching an absolute maximum of 1250 MWh at 19:00.Weekend consumption is lower and has less distinctive evening peaks.Source: §2.3 · p. 9
• Average specific consumption across the 200 profiles and four models is 22.2 kWh/100 km; median values are 22.7, 17.9, 15.7, and 14.5 kWh/100 km for Model 3, Kona, ID.3, and Zoe, respectively.Specific consumption is lowest in summer at 20.7 kWh/100 km and highest in winter at 23.6 kWh/100 km, with ambient temperature having a clear impact.Source: §2.3 · p. 9

2.4 Grid availability

• Charging-station availability reaches 90% between 3:00 and 5:00, with about 80% of vehicles connected at home and 10% on public streets; average availability falls to 70% around 11:00–12:00.During daytime, a relevant share of available stations is at workplaces; weekend availability is less peaky, with higher home and public-street shares during the day.Source: §2.4 · p. 10
• For one million BEV, median grid-connected power is 5.0–5.6 GW overnight and above 7 GW during daytime, reflecting the lower power rating of home charging.Home stations are rated at 3.6 kW, compared with 11 kW at workplaces and 22 kW in public areas.Source: §2.4 · p. 10

2.5 Grid electricity demand

• Median hourly demand peaks around 460 MWh under immediate full-capacity charging, while it rarely exceeds 400 MWh for immediate-balanced and 300 MWh for at-home-balanced charging.Balanced strategies charge vehicles below full power, reducing demand peaks and variance.Source: §2.5 · pp. 10–11
• At-home night-time balanced charging shows a working-day-night peak, with median hourly grid electricity demand of one million BEVs around 420 MWh, versus below 300 MWh at night between Friday evening and Monday morning because weekend vehicle use is lower.Shifting BEV charging to night-time would produce substantially less-smooth patterns than all-day charging.Source: §2.5 · p. 11

3 Discussion

• emobpy produces individual BEV profiles as four full-year, quarter-hourly series: mobility, driving electricity consumption, grid availability, and grid electricity demand.Source: §3 · pp. 11–12
• Users can choose the profile count and customize mobility data and assumptions, trading fleet diversity against computational burden in model applications.The examples use German mobility data and assumptions such as driver-type shares and charging-station availability and power ratings.Source: §3 · p. 12
• The profiles are intended as inputs to a wide range of energy, environmental, and economic analyses of electrified mobility.Source: §3 · p. 12
• Vehicle-focused modeling omits individual modal-split choice; adding it could relax the assumption that all trips are made with the same vehicle.Source: §3 · p. 12
• Mobility profiles draw on past behavior that may not reflect future behavior, though users can set alternative assumptions for future or counterfactual scenarios.Source: §3 · p. 12
• Profiles lack explicit spatial resolution, and drivers with service trips involving numerous work-related trips, such as taxi drivers, are excluded.The authors describe the spatial limitation as minor for models taking a macro perspective.Source: §3 · pp. 12–13
• The authors expect future collaborative development of the permissively licensed open-source tool could address these limitations.Source: §3 · p. 13

4 Methods

• Each BEV profile contains vehicle mobility, driving electricity consumption, grid availability, and grid electricity demand time series, built in that order.Mobility supplies locations and distances at each time step; later series use the preceding series as inputs.Source: §4 · p. 13
• A tour is a chronologically sorted list of trips, each represented by an origin-destination edge with departure time, distance, and duration attributes.The methods define an edge as a link connecting origin or destination nodes.Source: §4 · p. 13
• The sampling procedure draws discrete choices using their corresponding probabilities.For distance-duration relations with discrete ranges, it samples an integer distance uniformly within the range, then interpolates trip duration from the corresponding ranges.Source: §4 · p. 13

4.1 Vehicle mobility

• Tour generation uses driver-specific trip-count statistics, joint destination and departure-time distributions, and trip distance-duration relations.Week counts and a reference date can map days of the week, supporting weekday-specific statistics and temperature allocation for the driving-consumption series; commuter proportions and non-commuter inputs may come from data or assumptions.Source: §4.1 · p. 13
• For each calculation day, Select Tour samples trip counts by driver type and sequential destination-departure tuples from joint probabilities, avoiding repeated departure times.The remaining tuple probabilities are renormalized after each selection; sampled trips are then ordered chronologically.Source: §4.1 · p. 14
• Trip distance and duration are sampled from distance-duration statistics, and arrival and location times are derived from trip timing.The process creates origin-destination edges with departure time, distance, and duration as attributes.Source: §4.1 · p. 14
• A tour is retained only if it satisfies all plausibility rules, including minimum workplace time and whether its last trip heads home.Infeasible tours are discarded and the process repeats; the rules are listed in Table 5.Source: §4.1 · p. 14, Table 5 · p. 15

4.2 Driving electricity consumption

• Vehicle and passenger parameters, ambient temperature, and driving-cycle assumptions feed emobpy’s trip-consumption calculation from mobility time series.Vehicle inputs include BEV properties and customizable efficiencies and accessory power; the tool also provides temperature, vehicle-model, and cabin-insulation datasets.Source: §4.2 · p. 15

4.2.1 Custom driving cycles

• The default WLTC cycle is adapted by selecting the subclass closest to trip-average velocity, scaling it to that velocity, repeating it to trip duration, and deriving acceleration from velocity changes.Trip distance and duration determine average velocity; velocity and acceleration are simulated at high resolution for power and energy calculations.Source: §4.2.1 · p. 17

4.2.2 Vehicle tractive effort

• Tractive effort combines rolling resistance, aerodynamic drag, climbing force, and acceleration/inertia forces, with vehicle mass including passengers and rolling resistance depending on temperature and velocity.The inertial mass also depends on curb mass and gear ratio.Source: §4.2.2 · pp. 17–18

4.2.3 Motor power

• Wheel power is estimated at each time step, and motor output and input power are derived using transmission efficiency and a motor-efficiency function dependent on load fraction.Motor efficiency depends on angular speed and torque; the implemented general approach is described in and.Source: §4.2.3 · p. 18

4.2.4 Regenerative braking

• When tractive effort is negative, its absolute value defines regenerative braking power, whose recovery efficiency depends on acceleration and generator load fraction.Generator output power is estimated using generation efficiency, with the generator’s nominal capacity equal to the motor’s.Source: §4.2.4 · p. 19

4.2.5 Heating, cooling and accessories

• A cabin heat-balance model estimates heating or cooling demand from airflows, wall heat transfer, passenger heat, and device heat; radiation and latent heat are excluded.The device may be a resistor or heat pump, and accessory power is assumed constant.Source: §4.2.5 · p. 19
• The subsection specifies vehicle cabin-insulation configuration in Table 6, citing [27, 32, 34, 35].Source: Table 6 · p. 21

4.2.6 Energy consumption

• Total trip energy is defined by aggregating battery load over the trip’s duration, with discharge or regenerative-braking charging efficiency applied according to battery-load direction.Positive total motor, generator, auxiliary, and heating/cooling power draws energy from the battery; negative power corresponds to regenerative charging.Source: §4.2.6 · p. 21
• The required driving-consumption parameters are assumed identical across BEV models for simplicity and data availability.Table 7 provides the parameters used to estimate driving electricity consumption.Source: §4.2.6 · p. 21

4.3 Grid availability time series

• Grid-availability generation uses driving-consumption time series, battery size, charging efficiency, initial state of charge, and charging-station distributions and ratings.The imported time series includes time step, location, distance, and consumption; station types and power ratings are sampled for parking locations.Source: §4.3 · pp. 21–22
• Candidate station allocations are checked by calculating battery state of charge at every time step, adding grid energy while charging and subtracting driving consumption.The initial state of charge is exogenous, and feasibility is simulated with immediate full-capacity charging while connected.Source: §4.3 · p. 22
• If any state of charge falls outside 0–100%, emobpy reallocates charging stations; repeated unsuccessful allocations trigger a warning.The stated reasons include low station availability or power relative to trip lengths, or low battery capacity.Source: §4.3 · p. 23

• Grid availability, driving electricity consumption, vehicle locations, battery size, initial SoC, and charging efficiency inform time-step SoC and grid-drawn charging outputs.Source: §4.4 · pp. 23–24
• The two preset strategies are immediate - full capacity and immediate - balanced, alongside customizable charging strategies.The workflow imports input data before determining charging and battery SoC according to the selected strategy.Source: §4.4 · pp. 24, 26
• Immediate - full capacity charges at the full power rating whenever grid connection is available, aiming to reach 100% SoC as fast as possible, without foresight.Charging is reduced when less than the full rate is needed; during driving, charging is zero and motor consumption is subtracted from SoC.Source: §4.4 · p. 26
• Immediate - balanced charges when connected and uses perfect foresight across consecutive time steps parked at the same location to divide the energy needed for a full battery across parking time.It uses maximum station power only when 100% SoC cannot be reached before the next trip; otherwise, the charging rate is lower.Source: §4.4 · p. 26
• A customized strategy can specify charging hours and a parking location, such as 23–06 at home, and uses balanced charging.If SoC becomes negative, the model may charge outside the specified schedule and location boundary.Source: §4.4 · p. 26

• emobpy can be installed from the Python Package Index (PyPI).Source: Code availability statement · p. 26
• The code is provided under a permissive license in Zenodo.Source: Code availability statement · p. 26
• A script for generating the 200 BEV profiles in the case study is provided at GitLab.Source: Code availability statement · p. 26